Age-ing study: A prospective evaluation of geriatric, patient-reported and frailty assessments to predict outcomes in older adults with acute myeloid leukemia

Abstract

Abstract Background: Acute myeloid leukemia (AML) primarily affects older adults. Despite the survival benefit of intensive treatment, therapy decisions remain complex due to heterogeneity in physical function. Age alone does not adequately determine treatment tolerance or recovery potential. Tools including geriatric assessments (GAs), patient-reported outcomes (PROs), and frailty indices can predict outcomes but are underused in AML care. The AGE-ING (Assessment of Geriatric Evaluations Impact on New AML Guidance) study (NCT05909501) evaluated the utility of streamlined GA, PRO, and frailty tools to predict survival, toxicity, and functional outcomes in newly diagnosed patients with AML undergoing treatment. Interim results are presented here. Methods: This single-center, prospective observational study conducted at the University of Pennsylvania included patients newly diagnosed with AML aged ≥50, with Eastern Cooperative Oncology Group (ECOG) performance status (PS) 0–2 receiving intensive (daunorubicin/cytarabine) or non-intensive therapy (HMA/venetoclax). The objective was to evaluate whether pre-treatment GA and frailty assessments could predict overall survival (OS) at 30, 60 days, and 1 year (y) in these patients. Baseline GAs included GDS-15, Mini-Cog, PROs (physical function, depression, fatigue, anxiety, cognition), and the Short Physical Performance Battery (SPPB: balance, 4-meter [4m] walk, chair stand), repeated at days 14 and 30 after treatment. Frailty was assessed using the Deficit Accumulation Frailty Index (DAFI), based on pretreatment demographic, GA, and laboratory data and categorized as robust (<0.2), prefrail (0.2–0.35), or frail (>0.35). Multivariable Cox proportional hazards models and Kaplan–Meier (KM) models evaluated associations between baseline age, frailty, physical function, and OS, adjusted for treatment intensity. Results: As of June 2025, 47 of 100 patients were enrolled (21 [44.7%] in the intensive; 26 [55.3%] in the non-intensive treatment groups). Median age was 66 years (range 50–87); 39.6% aged ≥70. ECOG PS=0 in 42.9% (intensive) and 19.2% (non-intensive) treatment groups. European Leukemia Network 2022 risk classification in the intensive group was favorable/intermediate/adverse in 28.6%/42.8%/28.6% respectively, vs 15.4%/26.9%/57.7% in the non-intensive group. Overall patient frailty levels were classified as robust (68.1%), prefrail (27.7%) and frail (4.3%). At 1 month follow-up, SPPB scores declined for balance (mean score -0.6 [95% Confidence Interval (CI) 1.1 to -0.1], p=0.027), improved for 4m walk (mean score +0.4 [95% CI 0.0 to 0.8], p=0.036), and remained stable for chair stand (mean score ±0.0 [95% CI -0.4 to 0.4], p=0.998). PROs worsened significantly by 14 days in physical function (mean score -4.5 [95% CI -7.0 to -2.0], p<0.001), partially recovering at 1 month (mean score - 0.9 [95% CI -3.5 to 1.7], p=0.502). Fatigue (mean score -6.4 [95% CI -10.5 to -2.4], p=0.002), depression (mean score -5.4 [95% CI -7.7 to -3.0], p<0.001), and anxiety (mean score -6.6 [95% CI -9.4 to -3.8], p<0.001) worsened at 1 month. Mini-Cog scores improved at 1 month (mean score +0.6 [95% CI 0.1–1.0], p=0.009). At 30-days and 60-days, 1 (2.1%) and 2 patients (4.3%) respectively, had died, resulting in an overall incidence of 2.1 deaths per 100 person-months and hazard rate of death of 0.021 over the first 60 days. Interim KM estimates suggest patients ≥75 years old had a poorer OS at 1y compared with those <75 (hazard ratio [HR] 6.0, [95% CI 1.2 to 29.7], p=0.029). DAFI-defined frailty scores showed a trend toward inferior 1y OS (HR 2.3 [95% CI 1.0 to 5.7], p=0.063). Multivariate analysis showed no impact of frailty and treatment intensity on 1y OS. Estimated OS rates at 1y were 75.9% for robust patients receiving non-intensive treatment while prefrail/frail patients receiving non-intensive treatment were 56.6%. Numbers are too small for subgroup analysis and will be done with the final analysis. Conclusions: AGE-ING is the first study to assess serial measurements of GA, PROs, and frailty in patients with newly diagnosed AML in the real-world setting. Interim analysis shows patients ≥75 years and pre-frail/frail patients irrespective of age have worse OS. Final analysis is needed to evaluate differences based on treatment intensity, frailty scores, and change in measurements over time.